Paper Info
Title
Data Management for Automotive ECUs Based on Hybrid RAM-NVM Main Memory.
Abstract
More and more Electronic Control Units (ECUs) are applied into the automotive electrical and electronic systems. The system data and user data which ECUs need should be stored in Non-Volatile Memory (NVM) in order to avoid losing these data. However, NVM has the limitation on the numbers of writing so that the NVM is not suitable for the main memory [1]. In this paper, we propose a vehicle data allocation (VDA) algorithm that can reduce the writing times of NVM (if NVM is used to be main memory) or save random access memory (RAM) space (if RAM is used to be main memory) in automotive electronic control systems. In our method, both RAM and NVM are used as main memory. The data from automobile is classified into three categories and the automobile running phase is also divided into three phases. Our algorithm targets to manage the different categories data in NVM and RAM in different phases through the characteristics of the data. Our proposed algorithm can reduce the writing times in NVM effectively. Comparing with using NVM as the main memory, our method will reduce several hundred thousand writing times in NVM. Based on results of the case study, our algorithm can reduce on average 25% RAM space in the start phase, and even reduce on average 62.5% in most runtime.
Year
Venue
Field
2016
ICESS
Computer science,NVM Express,Real-time computing,Electronic systems,Data allocation,Control system,Data management,Random access,Embedded system,Automotive industry
DocType
Citations 
PageRank 
Conference
0
0.34
References 
Authors
0
7
Name
Order
Citations
PageRank
Junhuan Yang122.38
Jinyu Zhan286.23
Yiming Zhang314337.82
Wei Jiang4147.33
Lin Li532379.92
Li Zhu66522.91
Xuefei Tang700.68